How Elon's Twitter could look
You've probably heard: Elon Musk is trying to buy Twitter, with the rationale that it's a philanthropic move to preserve and promote democracy.
Whether or not he's successful, it's encouraged conversations about free speech, the role of social media and how a more transparent Twitter could theoretically work.
Undoubtedly a major asset in Elon's playbook will be the user experience. Not necessarily improving it from the typical lens of conversions, but rather with the goal of transparency and understanding.
Let's dive in:
1. Anti-viral notice
One of Musk's criticisms of the status quo, is that Twitter silently promotes or demotes content, through a mixture of both algorithmic and human decisions.
At some point, someone, for some reason, decided to promote Tweet X, and stop Tweet Y showing up in other people's feeds.
i.e., a notice could appear like this:
Clicking to learn more could—as we'll see in a moment—take you into the Matrix (Twitter's Matrix).
2. Open-sourced algorithm
In his recent Ted Talk, Elon outright mentioned releasing the algorithm(s) on GitHub—a popular tool for developers.
But whilst GitHub has an impressive 73 million users, Twitter has around 300 million.
To be clear, I'm not suggesting that we need 300 million people peer-reviewing the code, but rather everybody will need a basic understanding of how to control the content that they see.
As an example, why is this Tweet's reply appearing third, when there are plenty of other replies which have 100x more engagement, further down?
Now, consider an example like this, but where the Tweets are all politically charged. The lack of justification may obfuscate the real reason why that Tweet is being shown so highly—fuelling theories that it's an intentional act to help popularise a bias.
Putting aside the algorithmic peer-reviewing aspect, this poses a major UX challenge: how do you get people to interact with, understand and trust the mechanism in which content is served?
i.e., how do you get non-technical people to trust that there aren't controlling biases?
Clicking on the 'new' icon, on any Tweet could take you to a visual breakdown of the factors influencing the position and promotion of that Tweet.
This would allow non-technical users to peer-review the curation output with a greater understanding of the inputs.
This isn't the only solution, but it highlights the dilemma: it's not simply enough to fix the algorithm, you need to fix the public perception of the machine.
3. Better tools
Currently, when you report a Tweet as offensive, Twitter will "do something" with your complaint, and then give you two actions: to block or mute this user.
Hearing opinions that you don't agree with is a necessary result of free speech.
And as an organisation (or algorithm), being an arbiter of the rules is unfathomably hard. When is something disagreeable, and when does it encite violence?
As an example, if I'd reported one of Elon Musk's Tweets, Twitter could ask me why, and then help me identify and sever the connection that led to me being shown it in the first place.
i.e., by outlining the ingredients that led to that content being served.
Today, Twitter allows you to block or mute individuals, but makes very little effort to assist you in further curating the content that you see. It also doesn't help you understand why you're seeing Tweets from people that you don't follow.
As it stands, the user experience is not built as a curation tool, but optimised for endless consumption.
There's a similar UX challenge facing Musk, should the acquisition go ahead: helping users understand what's trending, and why.
I personally use Twitter's discovery tab daily—which I love because it's a real time reflection of society. Things trend on Twitter before they've broken on major news networks.
But, almost every time I use it, I see an item and wonder why I'm being shown that. It's very often neither notable, nor getting huge engagement.
Today, you can click, and mark hashtags as being uninteresting:
Which in a vague world of "tell us stuff you don't like, and we'll show you better stuff", feels akin to visiting a witch doctor, rather than traditional medicine.
Instead, the motivation to self-curate the trending content, relies on the basic understanding and confidence in how the topics are selected in the first place.
For example, when marking something as uninteresting, Twitter could help you select other topics that you're likely not interested in.
This isn't just a question of opening up the algorithm on GitHub, and letting developers pour over every line—but rather a long-term exercise of helping people utilise the mechanics of the game.
Humans are lazy, and many people will be more likely to moan endlessly about their messy feeds, rather than go into the settings and curate the content themselves.
Providing the functionality to do this is one thing, creating an experience which makes that process enjoyable is another.
So, to maintain the coherence of Twitter, without the vagueness of how it works, the complexities need to be approached with a thoughtful user experience in mind.